The present invention relates generally to metadata and more particularly to importing portions of metadata into a repository.
Typically, importing a metadata graph in one import would present an out of memory error due to the metadata graph being too large to load into a repository in a single execution. When importing metadata of a database or database schema into a repository, metadata can include tables, columns, and stored procedures, along with references between all the information. Loading the complete database or database schema into the repository does not allow for scalability. As a result, splitting the metadata into smaller portions and importing each portion of metadata individually avoids any out of memory errors. However, splitting the metadata into smaller portion results in a creation of multiple link points (i.e., a temporary representation of missing references between objects) in the metadata repository during the import process. As a result, the greater the number of link points results in a greater duration for importing the number of portions of metadata into the metadata repository.